Loss of cognition and dementia associated with neurological disease results from damage to neurons and synapses that serve as the anatomical substrata for memory, learning, and information processing. Despite much interest, biochemical pathways responsible for progressive neuronal loss in these disorders have not been elucidated.
Alzheimer's disease (AD) accounts for more than 15 million cases worldwide and is the most frequent cause of dementia in the elderly (Terry, R. D. et al. (eds.), ALZHEIMER'S DISEASE, Raven Press, New York, 1994). AD is thought to involve mechanisms which destroy neurons and synaptic connections. The neuropathology of this disorder includes formation of senile plaques which contain aggregates of Aβ1-42 (Selkoe, Neuron, 1991, 6:487-498; Yankner et al., New Eng. J. Med., 1991, 325:1849-1857; Price et al., Neurobiol. Aging, 1992, 13, 623-625; Younkin, Ann. Neurol., 1995, 37:287-288). Senile plaques found within the gray matter of AD patients are in contact with reactive microglia and are associated with neuron damage (Terry et al., “Structural Basis of the Cognitive Alterations in Alzheimer Disease”, ALZHEIMER'S DISEASE, NY, Raven Press, 1994, Ch. 11, 179-196; Terry, R. D. et al. (eds.), ALZHEIMER'S DISEASE, Raven Press, New York, 1994; Perlmutter et al., J. Neurosci. Res., 1992, 33:549-558). Plaque components from microglial interactions with Aβ plaques tested in vitro were found to stimulate microglia to release a potent neurotoxin, thus linking reactive microgliosis with AD neuronal pathology (Giulian et al., Neurochem. Int., 1995, 27:119-137).
Several lines of evidence now support the concept that microglia-derived neurotoxins contribute to AD pathology. First, microglia-derived toxins can be extracted from AD brain regions laden with plaques but not from identical brain regions in age-matched control or ALS brain tissues (Giulian et al. (1995) Neurochem. Int., 27: 119-137; Giulian et al. (1996) J. Neurosci., 16: 6021-6037). Second, regional distributions of toxic activity show the greatest concentrations of microglia-derived neuron poisons in neocortical tissues and hippocampi of AD (vs. controls or ALS), areas containing large numbers of reactive microglia. In contrast, cerebellum, white matter, and neocortical tissues from normal or ALS patients, which had few, if any, reactive microglial clusters, show little neurotoxic activity. Moreover, the relative number of reactive microglial clusters in each brain region is significantly correlated to the level of neurotoxic activity extracted from that region (p<0.005). Third, isolated plaque fragments or synthetic human Aβ1-40 or Aβ1-42 peptides are found to activate human microglia to release neurotoxins in culture (Giulian et al. (1995) Neurochem. Int., 27: 119-137; Giulian et al. (1996) J. Neurosci., 16: 6021-6037). No neurotoxic effects, however, are detected when plaques or peptides were placed directly atop neurons or when microglia are exposed to fractions lacking plaques isolated from AD, ALS, or normal, aged control brains (Giulian et al. (1995) Neuroclem. Int. 27: 119-137; Giulian et al. (1996) J. Neurosci. 16: 6021-6037). Thus, the toxic effects of isolated plaques on neurons are indirect and mediated by a neurotoxic activity released from plaque-stimulated microglia. Fourth, there is neurotoxic activity found in CSF from AD patients, but not detected in samples from disease controls (U.S. Pat. No. 6,043,283 to Giulian; Giulian et al. (1999) Am. J. Hum. Genet., 65:13-18). Fifth, infusion of Aβ-coupled microspheres into hippocampus produces inflammatory responses at the site of infusion in rats (U.S. Pat. No. 6,043,283 to Giulian). Together, these data indicate that plaque-activation of microglia through contact with Aβ peptides produces neuron-killing factors in discrete areas of AD brain (Giulian et al. (1995) Neurochem. Int., 27: 119-137).
Although most patients developing AD will go through a transient period of mild cognitive impairment (MCI), they will often not present to a physician during this early phase of the disease. There is a consensus among research groups that subjects with MCI are at increased risk for progressing to AD (Grundman et al. (1996) Neurology 46:403; Flicker et al. (1991) Neurology 41:1006-1009; Masur et al. (1994) Neurology 44:1427-1432; Tierney et al. (1996) Neurology 46:149-154). Memory impairment is commonly the most prominent feature of MCI but might include other patterns including defects primarily in language or visuomotor performance (Hughes et al. (1982) Br. J. Psychiatry, 140:566-572; Berg (1988) Psychopharmacol. Bull., 24:637-639; Morris (1993) Neurology, 43:2412-2414; Rubin et al. (1989) Arch. Neurol., 46:379-382; Grundman et al. (1996) Neurology, 46:403; Flicker et al. (1991) Neurology, 41:1006-1009; Masur et al. (1994) Neurology, 44:1427-1432; Tierney et al. (1996) Neurology, 46:149-154). Attempts at characterizing mild cognitive impairment have been carried out using the Clinical Dementia Rating (CDR) Scale, which rates the severity of dementia as absent, mild, moderate, or severe. Rubin et al. ((1989) Arch. Neurol., 46:379-382) concluded that individuals with a CDR of 0.5 likely have “very mild” AD in the majority of cases [The CDR 0.5 classification is characterized by consistent forgetfulness, which is mild with little if any impairment in other functions such as orientation, community affairs, home, and hobbies, judgment, and personal care.] Other measures also have been used to identify MCI subjects. For example, poor delayed recall has been shown to be the best predictor of progression, the best predictor of subsequent dementia in non demented elderly subjects, and the best discriminator between normal aging and mild AD (Flicker et al. (1991) Neurology, 41:1006-1009, Masur et al. (1994) Neurology, 44:1427-1432; Tierney et al. (1996) Neurology, 46:149-154). The time required for subjects with MCI to develop a clinical diagnosis of AD has been estimated by the Alzheimer's Disease Cooperative Study (ADCS) at about 30% at 2 years and 45% at 3 years.
HIV-1 infection and neuro-AIDS produce devastating effects upon the brain and spinal cord. Although the underlying anatomical basis for impaired cognition during HIV-1 infection remains obscure, there is a reduction of up to 40% of large neurons scattered throughout the neocortex in advanced disease with dementia (Masliah et al. (1992) J. Neuropath Exp Neurol., 51: 585-593) and a striking early loss of synapses (Asare et al. (1996) Am J Path 148: 31-38; Everall et al. (1993) J. Neuropath. Exp. Neurol. 52: 561-566).
HIV-1 associated dementia (HAD) is characterized by cognitive dysfunction, declining motor performance, and behavioral changes. It occurs primarily in the more advanced stages of HIV infection when CD4 cell counts are relatively low. While the progression of dysfunction is variable, it is regarded as a serious complication with fatal outcome. The diagnosis of cognitive loss due to HIV is by process of exclusion—no approved marker exists to monitor HIV-specific injury to the CNS. Without such a marker, there are no clinical indications to evaluate patients until significant functional loss appears and there are few opportunities to develop new treatment strategies to prevent HIV brain damage. Therefore, it is very desirable to identify patients at early pre-symptomatic stages.
Prior to HAART (defined here as combination therapy using 3 or more anti-retroviral agents), 60% of those with AIDS developed dementia. This incidence appears to have fallen to about 10 to 15%, but cognitive dysfunction remains a problem for over half of the HIV/AIDS population (Giulian et al. (1990) Science, 250: 1593-1596; Giulian et al. (1993) Proc. Natl. Acad. Sci., 90:2769-2773; Giulian (1995) In: NEUROGLIA (H Kettenmann, B Ransom Eds) Oxford University Press, pp. 671-684; Giulian et al. (1998) In: INFLAMMATORY MECHANISMS OF NEURODGENERATION AND ITS MANAGEMENT (P. Wood, ed.); Humana Press, Vol 4, pp. 109-128).
HIV-1 brain pathology involves diffuse synaptic damage in the neocortex, the loss of cortical neurons, and a population of infected, reactive mononuclear phagocytes, including invading blood monocytes, microglia, and multi-nucleated giant cells. These giant cells represent a fusion of HIV-infected mononuclear phagocytes that are coated with gp120, the retroviral envelope protein; presence of giant cells has been correlated with cognitive impairment during HIV-1 infection. Currently, most research groups in the field agree that poisons released by infected mononuclear phagocytes are a primary cause of cognitive loss in the HIV-1(+) population (Vitokovic et al. (1998) Medical Sciences, 321: 1015-1021; Morgello et al. (2001) Neuropath. App. Neurobiol., 27: 326-335; Lawrence et al. (2002) Microbes and Infection, 4: 301-308; Masliah et al. (1992) J. Neuropath. Exp. Neurol., 51: 585-593; Maslliah et al. (1995) J. Neuropath. and Exp. Neurol., 54: 350-357; Asare et al. (1996) Am. J. Path., 148: 31-38; Everall et al. (1993) J. Neuropath. Exp. Neurol., 52: 561-566).
Several lines of evidence now support the concept that mononuclear phagocyte-derived neurotoxins contribute to the neuron injury within brain during HIV-1 infection. First, HIV-1 neither infected neurons nor showed a direct toxic effect upon neurons (Giulian et al. (1996) J. Neurosci., 16:3139-3153, Giulian et al. (1990) Science 250: 1593-1596; Levine et al. (1976) Biochim. Biophys. Acta, 452: 458467). Second, HIV-1 mononuclear phagocytes (THP-1, U937, human blood monocytes, and human brain microglia) released neurotoxins when infected in vitro with HIV-1; in contrast, lymphocytes (H9, human blood lymphocytes) did not (Giulian et al. (1996) J. Neurosci., 16:3139-3153; Giulian et al. (1990) Science, 250: 1593-1596). Third, human mononuclear phagocytes (blood monocytes and microglia) isolated from infected donors released the same neurotoxin as recovered from in vitro experiments; again, isolated infected lymphocytes did not (Giulian et al. (1996) J. Neurosci., 16:3139-3153). Fourth, neurotoxic activity can be recovered from brain tissues of infected individuals (Giulian et al. (1993) Proc. Natl. Acad. Sci., 90:2769-2773; Giulian (1995) In: NEUROGLIA (H Kettenmann, B Ransom, Eds,) Oxford University Press, pp. 671-684; Giulian et al. (1998) In: INFLAMMATORY MECHANISMS OF NEURODEGENERATION AND ITS MANAGEMENT (P. Wood, ed.); Humana Press, Vol 4, pp. 109-128). Fifth, gp120, the viral envelope glycoprotein, can stimulate neurotoxin release from human blood monocytes and microglia; other viral proteins including tat did not (Levine et al. (1976) Biochim. Biophys. Acta, 452: 458-467). Sixth, high concentrations of neurotoxin were found in the cerebrospinal fluid of HIV-1(+) individuals. And seventh, a family of neurotoxic heparan oligosaccharides can be isolated from HIV-1 infected cells and from HIV CSF.
Although reactive mononuclear phagocytes release a number of bio-active substances, few of these compounds are actually able to harm neurons at concentrations found to exist in neurodegenerative disease (Hardy et al. (2002) Science, 297:353; Mourdian et al., (1989) Neurosci. Lett., 105: 233; Milstein et al. (1994) J. Neurochemistry, 63, 1178; Giulian et al. (1990) Science, 250:1593). Moreover, few of such candidate neuron poisons are present in both AD and HAD. For example, increased tissue concentrations of “toxic” forms of Aβ1-42 are characteristic for AD (Hardy et al. (2002) Science, 297:353), but do not occur in HAD. Similarly, elevated quinolinic acid levels occur in the cerebrospinal fluid (CSF) of subjects with HAD (Mourdian et al. (1989) Neurosci. Lett., 105:233), but not in those with AD (Milstein, et al. (1994) J. Neurochemistry, 63: 1178). In contrast, both AD and HAD brain tissues contain a heterogeneous group of small stable molecules with potent neurotoxic actions (Giulian et al. (1990) Science, 250:1593; Giulian et al. (1995) Neurochem. Int., 27:119; Giulian et al. (1996) J. Neuroscience 16: 6021). Cultured mononuclear phagocytes activated by exposure to amyloid plaques, synthetic β-amyloid peptides, HIV-1, or gp120, produce these same neurotoxins (Giulian, et al. (1993) Proc. Natl. Acad. Sci. USA, 90: 2769; Giulian et al. (1998) J. Biol. Chem., 273: 29719). Such observations suggest that a common, though unidentified, pathway mediates immune-driven neuron pathology in both AD and HAD.
As the clinical expression of neurological disease may occur only after a significant degree of neuron loss and synaptic damage beyond a critical threshold necessary for adequate adaptive function, early pre-symptomatic detection of disease pathology offers the opportunity to slow disease progression. The present invention provides methods for diagnosis of neurological disease and risk for loss of cognition, including, for example, Alzheimer's disease, HIV-1 associated dementia (HAD), neuro-AIDS, Creutzfeldt-Jakob disease, Mild Cognitive Impairment (MCI), prion disease, minor cognitive/motor dysfunction, acute stroke, or acute trauma. The methods of the invention allow early detection of neurological disease and risk for loss of cognition, thereby allowing earlier intervention in the progression of disease. Also provided are methods for monitoring the progression and treatment of neurological disease by monitoring encephalotoxin levels in a subject.